Reproduce “TinyCardioUNet: IMU-to-ECG Translation with Graph-Encoded Inter-Axis Dependencies and Tensor Decomposition-Based Parameter Reduction” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Seungwoo Han, Ingon Chanpornpakdi, Motoi Noda, Puwadej Leelasiri, Ibuki Hiruma, Toshihisa Tanaka
TinyCardioUNet reconstructs ECG-like waveforms from all six chest-worn inertial-sensor axes, using a U-Net with graph-based information exchange at its bottleneck and post-training tensor decomposition. After 10-epoch fine-tuning, it has 36,004 parameters and reports test-segment RMSE 0.098 ± 0.047 and waveform correlation 0.677 ± 0.177. Evaluation covered 29 healthy male volunteers and controlled Gaussian noise, not real-world motion.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.